Project3
STATS 220 Semester One 2024
Yuekun Yao
2024-04-24
Introduction
I chose starry sky as my two search words in the website of pixabay, which are the pictures I love to watching and are beautiful.
- photo1 features: The tags are
Astronomy,Night sky,Cosmos image. There are 94,488 views, 78,253 downloads and 199 likes. The photo was shot by Camera Nikon D750. - photo2 features: The tags are
Heaven,Stars,Fog image. There are 29,053 views,19,016 downloads and 134 likes. The photo was shot by Camera Fujifilm X-T2. - photo3 features: The tags are
Lake,Water,Stars image. There are 27,212 views,18,650 downloads and 123 likes. The photo was shot by Camera Fujifilm X-T2.
- photo1 features: The tags are
The page URL of selected photos
photo_data %>%
select(pageURL) %>%
knitr::kable()
Key features of selected photos
- The average downloads of selected photos is 275217.01.
- The median downloads of selected photos is 42129.
- The sum of comments of selected photos is 6351.
- The average number of views per photo is 407326.09.
- There are 4 photos which was gotten more a million of downloads among selected 67 photos.
#Calculate the amount of photos which is more million downloads among the selected photos.
amount_more_million_downloads <- photo_data %>%
group_by(more_million_downloads) %>%
summarise(amount = n())
amount_more_million_downloads## # A tibble: 2 × 2
## more_million_downloads amount
## <chr> <int>
## 1 no 63
## 2 yes 4Creativity
Create a plot of the ratio of downloads to views for each user by
using group_by() and summarise() function and
using the summary value into ggplot() and
geom_point() function.
#Calculate the ratio of between downloads and views for each users
user_download_ratio <- photo_data %>%
group_by(user_id) %>%
summarise(downloads_ratio = sum(downloads)/sum(views))
#Create a plot of the ratio of downloads to views for each user
user_download_ratio_plot <- user_download_ratio %>%
ggplot() +
geom_point(aes(x=user_id, y = downloads_ratio)) +
labs(
title = "The downloads ratio of users",
subtitle = "The ratio of downloads to views for each user",
caption = "Source: Pixabay",
x = "User ID",
y = "Downloads Ratio"
)
options(scipen = 999)
user_download_ratio_plot
Learning reflection
From the Module 3:Creating new variables and data tables, I learned
to identify key features of JSON, how to write simple queries using SQL,
how to manipulate data frames using {dplyr}, how to source data from an
API, how to report on key features of data, how to use {dplyr} to
produce summaries, how to using R & RMarkdown for reproducible
reporting including drop_na(), na.omit(),
inline R code & knitr::kable(). And there
are some important functions I learned in module 3, such as
fromJSON(),mutate(),select(),group_by(),summarise(),geom_point().
In the future, I want to learn more about how to analyse the data using some packages on the R.
Appendix
library(magick)
library(tidyverse)
library(jsonlite)
#Explore the code!
json_data <- fromJSON("pixabay_data.json")
pixabay_photo_data <- json_data$hits
selected_photos <- pixabay_photo_data %>%
select(previewURL,pageURL,downloads,tags, likes, user_id, views, comments) %>%
mutate(tags_length = str_count(tags, "\\S+")) %>%
mutate(more_million_downloads = ifelse(downloads > 1000000, "yes", "no")) %>%
mutate(if_contain_search_words = ifelse(str_detect(str_to_lower(tags),
"starry|sky"),
"yes",
"no")) %>%
filter(more_million_downloads == "yes" | if_contain_search_words == "yes" )
write_csv(selected_photos, "selected_photos.csv")
#Calculate the mean downloads of selected photos
mean_downloads <- selected_photos$downloads %>%
mean()
mean_downloads
#Calculate the mean tags_length of selected photos
mean_tags_length <- selected_photos$tags_length %>%
mean()
mean_tags_length
#Calculate the sum of comments of selected photos
sum_comments <- selected_photos$comments %>%
sum()
sum_comments
#Calculate the sum of more million downloads and less million downloads
sum_more_million_downloads <- selected_photos %>%
group_by(more_million_downloads) %>%
summarise(sum_downloads = sum(downloads))
sum_more_million_downloads
# Calculate the mean likes of different users
user_likes <- selected_photos %>%
group_by(user_id) %>%
summarise(mean_like = mean(likes))
user_likes
#Calculate the amount of photos which is more million downloads among the selected photos.
amount_more_million_downloads <- photo_data %>%
group_by(more_million_downloads) %>%
summarise(amount = n())
amount_more_million_downloads
#Calculate the ratio of between downloads and views for each users
user_download_ratio <- selected_photos %>%
group_by(user_id) %>%
summarise(downloads_ratio = sum(downloads)/sum(views))
user_download_ratio
#Create an animated GIF of the photos in your selected_photos data frame
img_urls <- selected_photos$previewURL %>% na.omit()
img_urls %>%
image_read() %>%
image_scale(200) %>%
image_animate(fps = 2) %>%
image_write("my_photos.gif")
#Create a plot of the ratio of downloads to views for each user
user_download_ratio_plot <- user_download_ratio %>%
ggplot() +
geom_point(aes(x=user_id, y = downloads_ratio)) +
labs(
title = "The downloads ratio of users",
subtitle = "The ratio of downloads to views for each user",
caption = "Source: Pixabay",
x = "User ID",
y = "Downloads Ratio"
)
options(scipen = 999)
user_download_ratio_plot